linux-stable.git/drivers/hid, branch v6.6.2

HID: logitech-hidpp: Move get_wireless_feature_index() check to hidpp_connect_event()

2023-11-20T10:59:22+00:00

[ Upstream commit ba9de350509504fb748837b71e23d7e84c83d93c ]

Calling get_wireless_feature_index() from probe() causes
the wireless_feature_index to only get set for unifying devices which
are already connected at probe() time. It does not get set for devices
which connect later.

Fix this by moving get_wireless_feature_index() to hidpp_connect_event(),
this does not make a difference for devices connected at probe() since
probe() will queue the hidpp_connect_event() for those at probe time.

This series has been tested on the following devices:
Logitech Bluetooth Laser Travel Mouse (bluetooth, HID++ 1.0)
Logitech M720 Triathlon (bluetooth, HID++ 4.5)
Logitech M720 Triathlon (unifying, HID++ 4.5)
Logitech K400 Pro (unifying, HID++ 4.1)
Logitech K270 (eQUAD nano Lite, HID++ 2.0)
Logitech M185 (eQUAD nano Lite, HID++ 4.5)
Logitech LX501 keyboard (27 Mhz, HID++ builtin scroll-wheel, HID++ 1.0)
Logitech M-RAZ105 mouse (27 Mhz, HID++ extra mouse buttons, HID++ 1.0)

And by bentiss:
Logitech Touchpad T650 (unifying)
Logitech Touchpad T651 (bluetooth)
Logitech MX Master 3B (BLE)
Logitech G403 (plain USB / Gaming receiver)

Fixes: 0da0a63b7cba ("HID: logitech-hidpp: Support WirelessDeviceStatus connect events")
Signed-off-by: Hans de Goede 
Link: https://lore.kernel.org/r/20231010102029.111003-4-hdegoede@redhat.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

HID: logitech-hidpp: Revert "Don't restart communication if not necessary"

2023-11-20T10:59:22+00:00

[ Upstream commit 55bf70362ffc4ddd7c8745e2fe880edac00e4aff ]

Commit 91cf9a98ae41 ("HID: logitech-hidpp: make .probe usbhid capable")
makes hidpp_probe() first call hid_hw_start(hdev, 0) to allow IO
without connecting any hid subdrivers (hid-input, hidraw).

This is done to allow to retrieve the device's name and serial number
and store these in hdev->name and hdev->uniq.

Then later on IO was stopped and started again with hid_hw_start(hdev,
HID_CONNECT_DEFAULT) connecting hid-input and hidraw after the name
and serial number have been setup.

Commit 498ba2069035 ("HID: logitech-hidpp: Don't restart communication
if not necessary") changed the probe() code to only do the start with
a 0 connect-mask + restart later for unifying devices.

But for non unifying devices hdev->name and hdev->uniq are updated too.
So this change re-introduces the problem for which the start with
a 0 connect-mask + restart later behavior was introduced.

The previous patch in this series changes the unifying path to instead of
restarting IO only call hid_connect() later. This avoids possible issues
with restarting IO seen on non unifying devices.

Revert the change to limit the restart behavior to unifying devices to
fix hdev->name changing after userspace facing devices have already been
registered.

This series has been tested on the following devices:
Logitech Bluetooth Laser Travel Mouse (bluetooth, HID++ 1.0)
Logitech M720 Triathlon (bluetooth, HID++ 4.5)
Logitech M720 Triathlon (unifying, HID++ 4.5)
Logitech K400 Pro (unifying, HID++ 4.1)
Logitech K270 (eQUAD nano Lite, HID++ 2.0)
Logitech M185 (eQUAD nano Lite, HID++ 4.5)
Logitech LX501 keyboard (27 Mhz, HID++ builtin scroll-wheel, HID++ 1.0)
Logitech M-RAZ105 mouse (27 Mhz, HID++ extra mouse buttons, HID++ 1.0)

And by bentiss:
Logitech Touchpad T650 (unifying)
Logitech Touchpad T651 (bluetooth)
Logitech MX Master 3B (BLE)
Logitech G403 (plain USB / Gaming receiver)

Fixes: 498ba2069035 ("HID: logitech-hidpp: Don't restart communication if not necessary")
Signed-off-by: Hans de Goede 
Link: https://lore.kernel.org/r/20231010102029.111003-3-hdegoede@redhat.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

HID: logitech-hidpp: Don't restart IO, instead defer hid_connect() only

2023-11-20T10:59:22+00:00

[ Upstream commit 11ca0322a41920df2b462d2e45b0731e47ff475b ]

Restarting IO causes 2 problems:

1. Some devices do not like IO being restarted this was addressed in
   commit 498ba2069035 ("HID: logitech-hidpp: Don't restart communication
   if not necessary"), but that change has issues of its own and needs to
   be reverted.

2. Restarting IO and specifically calling hid_device_io_stop() causes
   received packets to be missed, which may cause connect-events to
   get missed.

Restarting IO was introduced in commit 91cf9a98ae41 ("HID: logitech-hidpp:
make .probe usbhid capable") to allow to retrieve the device's name and
serial number and store these in hdev->name and hdev->uniq before
connecting any hid subdrivers (hid-input, hidraw) exporting this info
to userspace.

But this does not require restarting IO, this merely requires deferring
calling hid_connect(). Calling hid_hw_start() with a connect-mask of
0 makes it skip calling hid_connect(), so hidpp_probe() can simply call
hid_connect() later without needing to restart IO.

Remove the stop + restart of IO and instead just call hid_connect() later
to avoid the issues caused by restarting IO.

Now that IO is no longer stopped, hid_hw_close() must be called at the end
of probe() to balance the hid_hw_open() done at the beginning probe().

This series has been tested on the following devices:
Logitech Bluetooth Laser Travel Mouse (bluetooth, HID++ 1.0)
Logitech M720 Triathlon (bluetooth, HID++ 4.5)
Logitech M720 Triathlon (unifying, HID++ 4.5)
Logitech K400 Pro (unifying, HID++ 4.1)
Logitech K270 (eQUAD nano Lite, HID++ 2.0)
Logitech M185 (eQUAD nano Lite, HID++ 4.5)
Logitech LX501 keyboard (27 Mhz, HID++ builtin scroll-wheel, HID++ 1.0)
Logitech M-RAZ105 mouse (27 Mhz, HID++ extra mouse buttons, HID++ 1.0)

And by bentiss:
Logitech Touchpad T650 (unifying)
Logitech Touchpad T651 (bluetooth)
Logitech MX Master 3B (BLE)
Logitech G403 (plain USB / Gaming receiver)

Fixes: 498ba2069035 ("HID: logitech-hidpp: Don't restart communication if not necessary")
Suggested-by: Benjamin Tissoires 
Signed-off-by: Hans de Goede 
Link: https://lore.kernel.org/r/20231010102029.111003-2-hdegoede@redhat.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

HID: uclogic: Fix a work->entry not empty bug in __queue_work()

2023-11-20T10:59:22+00:00

[ Upstream commit d45f72b3c275101a620dd69881343e0dda72f877 ]

When CONFIG_HID_UCLOGIC=y and CONFIG_KUNIT_ALL_TESTS=y, launch
kernel and then the below work->entry not empty bug occurs.

In hid_test_uclogic_exec_event_hook_test(), the filter->work is not
initialized to be added to p.event_hooks->list, and then the
schedule_work() in uclogic_exec_event_hook() will call __queue_work(),
which check whether the work->entry is empty and cause the below
warning call trace.

So call INIT_WORK() with a fake work to solve the issue. After applying
this patch, the below work->entry not empty bug never occurs.

 WARNING: CPU: 0 PID: 2177 at kernel/workqueue.c:1787 __queue_work.part.0+0x780/0xad0
 Modules linked in:
 CPU: 0 PID: 2177 Comm: kunit_try_catch Tainted: G    B   W        N 6.6.0-rc2+ #30
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
 RIP: 0010:__queue_work.part.0+0x780/0xad0
 Code: 44 24 20 0f b6 00 84 c0 74 08 3c 03 0f 8e 52 03 00 00 f6 83 00 01 00 00 02 74 6f 4c 89 ef e8 c7 d8 f1 02 f3 90 e9 e5 f8 ff ff <0f> 0b e9 63 fc ff ff 89 e9 49 8d 57 68 4c 89 e6 4c 89 ff 83 c9 02
 RSP: 0000:ffff888102bb7ce8 EFLAGS: 00010086
 RAX: 0000000000000000 RBX: ffff888106b8e460 RCX: ffffffff84141cc7
 RDX: 1ffff11020d71c8c RSI: 0000000000000004 RDI: ffff8881001d0118
 RBP: dffffc0000000000 R08: 0000000000000001 R09: ffffed1020576f92
 R10: 0000000000000003 R11: ffff888102bb7980 R12: ffff888106b8e458
 R13: ffff888119c38800 R14: 0000000000000000 R15: ffff8881001d0100
 FS:  0000000000000000(0000) GS:ffff888119c00000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: ffff888119506000 CR3: 0000000005286001 CR4: 0000000000770ef0
 DR0: ffffffff8fdd6ce0 DR1: ffffffff8fdd6ce1 DR2: ffffffff8fdd6ce3
 DR3: ffffffff8fdd6ce5 DR6: 00000000fffe0ff0 DR7: 0000000000000600
 PKRU: 55555554
 Call Trace:
  
  ? __warn+0xc9/0x260
  ? __queue_work.part.0+0x780/0xad0
  ? report_bug+0x345/0x400
  ? handle_bug+0x3c/0x70
  ? exc_invalid_op+0x14/0x40
  ? asm_exc_invalid_op+0x16/0x20
  ? _raw_spin_lock+0x87/0xe0
  ? __queue_work.part.0+0x780/0xad0
  ? __queue_work.part.0+0x249/0xad0
  queue_work_on+0x48/0x50
  uclogic_exec_event_hook.isra.0+0xf7/0x160
  hid_test_uclogic_exec_event_hook_test+0x2f1/0x5d0
  ? try_to_wake_up+0x151/0x13e0
  ? uclogic_exec_event_hook.isra.0+0x160/0x160
  ? _raw_spin_lock_irqsave+0x8d/0xe0
  ? __sched_text_end+0xa/0xa
  ? __sched_text_end+0xa/0xa
  ? migrate_enable+0x260/0x260
  ? kunit_try_run_case_cleanup+0xe0/0xe0
  kunit_generic_run_threadfn_adapter+0x4a/0x90
  ? kunit_try_catch_throw+0x80/0x80
  kthread+0x2b5/0x380
  ? kthread_complete_and_exit+0x20/0x20
  ret_from_fork+0x2d/0x70
  ? kthread_complete_and_exit+0x20/0x20
  ret_from_fork_asm+0x11/0x20
  

Fixes: a251d6576d2a ("HID: uclogic: Handle wireless device reconnection")
Signed-off-by: Jinjie Ruan 
Reviewed-by: José Expósito 
Link: https://lore.kernel.org/r/20231009064245.3573397-3-ruanjinjie@huawei.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

HID: uclogic: Fix user-memory-access bug in uclogic_params_ugee_v2_init_event_hooks()

2023-11-20T10:59:22+00:00

[ Upstream commit 91cfe0bbaa1c434d4271eb6e1d7aaa1fe8d121f6 ]

When CONFIG_HID_UCLOGIC=y and CONFIG_KUNIT_ALL_TESTS=y, launch kernel and
then the below user-memory-access bug occurs.

In hid_test_uclogic_params_cleanup_event_hooks(),it call
uclogic_params_ugee_v2_init_event_hooks() with the first arg=NULL, so
when it calls uclogic_params_ugee_v2_has_battery(), the hid_get_drvdata()
will access hdev->dev with hdev=NULL, which will cause below
user-memory-access.

So add a fake_device with quirks member and call hid_set_drvdata()
to assign hdev->dev->driver_data which avoids the null-ptr-def bug
for drvdata->quirks in uclogic_params_ugee_v2_has_battery(). After applying
this patch, the below user-memory-access bug never occurs.

 general protection fault, probably for non-canonical address 0xdffffc0000000329: 0000 [#1] PREEMPT SMP KASAN
 KASAN: probably user-memory-access in range [0x0000000000001948-0x000000000000194f]
 CPU: 5 PID: 2189 Comm: kunit_try_catch Tainted: G    B   W        N 6.6.0-rc2+ #30
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
 RIP: 0010:uclogic_params_ugee_v2_init_event_hooks+0x87/0x600
 Code: f3 f3 65 48 8b 14 25 28 00 00 00 48 89 54 24 60 31 d2 48 89 fa c7 44 24 30 00 00 00 00 48 c7 44 24 28 02 f8 02 01 48 c1 ea 03 <80> 3c 02 00 0f 85 2c 04 00 00 48 8b 9d 48 19 00 00 48 b8 00 00 00
 RSP: 0000:ffff88810679fc88 EFLAGS: 00010202
 RAX: dffffc0000000000 RBX: 0000000000000004 RCX: 0000000000000000
 RDX: 0000000000000329 RSI: ffff88810679fd88 RDI: 0000000000001948
 RBP: 0000000000000000 R08: 0000000000000000 R09: ffffed1020f639f0
 R10: ffff888107b1cf87 R11: 0000000000000400 R12: 1ffff11020cf3f92
 R13: ffff88810679fd88 R14: ffff888100b97b08 R15: ffff8881030bb080
 FS:  0000000000000000(0000) GS:ffff888119e80000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 0000000000000000 CR3: 0000000005286001 CR4: 0000000000770ee0
 DR0: ffffffff8fdd6cf4 DR1: ffffffff8fdd6cf5 DR2: ffffffff8fdd6cf6
 DR3: ffffffff8fdd6cf7 DR6: 00000000fffe0ff0 DR7: 0000000000000600
 PKRU: 55555554
 Call Trace:
  
  ? die_addr+0x3d/0xa0
  ? exc_general_protection+0x144/0x220
  ? asm_exc_general_protection+0x22/0x30
  ? uclogic_params_ugee_v2_init_event_hooks+0x87/0x600
  ? sched_clock_cpu+0x69/0x550
  ? uclogic_parse_ugee_v2_desc_gen_params+0x70/0x70
  ? load_balance+0x2950/0x2950
  ? rcu_trc_cmpxchg_need_qs+0x67/0xa0
  hid_test_uclogic_params_cleanup_event_hooks+0x9e/0x1a0
  ? uclogic_params_ugee_v2_init_event_hooks+0x600/0x600
  ? __switch_to+0x5cf/0xe60
  ? migrate_enable+0x260/0x260
  ? __kthread_parkme+0x83/0x150
  ? kunit_try_run_case_cleanup+0xe0/0xe0
  kunit_generic_run_threadfn_adapter+0x4a/0x90
  ? kunit_try_catch_throw+0x80/0x80
  kthread+0x2b5/0x380
  ? kthread_complete_and_exit+0x20/0x20
  ret_from_fork+0x2d/0x70
  ? kthread_complete_and_exit+0x20/0x20
  ret_from_fork_asm+0x11/0x20
  
 Modules linked in:
 Dumping ftrace buffer:
    (ftrace buffer empty)
 ---[ end trace 0000000000000000 ]---
 RIP: 0010:uclogic_params_ugee_v2_init_event_hooks+0x87/0x600
 Code: f3 f3 65 48 8b 14 25 28 00 00 00 48 89 54 24 60 31 d2 48 89 fa c7 44 24 30 00 00 00 00 48 c7 44 24 28 02 f8 02 01 48 c1 ea 03 <80> 3c 02 00 0f 85 2c 04 00 00 48 8b 9d 48 19 00 00 48 b8 00 00 00
 RSP: 0000:ffff88810679fc88 EFLAGS: 00010202
 RAX: dffffc0000000000 RBX: 0000000000000004 RCX: 0000000000000000
 RDX: 0000000000000329 RSI: ffff88810679fd88 RDI: 0000000000001948
 RBP: 0000000000000000 R08: 0000000000000000 R09: ffffed1020f639f0
 R10: ffff888107b1cf87 R11: 0000000000000400 R12: 1ffff11020cf3f92
 R13: ffff88810679fd88 R14: ffff888100b97b08 R15: ffff8881030bb080
 FS:  0000000000000000(0000) GS:ffff888119e80000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 0000000000000000 CR3: 0000000005286001 CR4: 0000000000770ee0
 DR0: ffffffff8fdd6cf4 DR1: ffffffff8fdd6cf5 DR2: ffffffff8fdd6cf6
 DR3: ffffffff8fdd6cf7 DR6: 00000000fffe0ff0 DR7: 0000000000000600
 PKRU: 55555554
 Kernel panic - not syncing: Fatal exception
 Dumping ftrace buffer:
    (ftrace buffer empty)
 Kernel Offset: disabled
 Rebooting in 1 seconds..

Fixes: a251d6576d2a ("HID: uclogic: Handle wireless device reconnection")
Signed-off-by: Jinjie Ruan 
Reviewed-by: José Expósito 
Link: https://lore.kernel.org/r/20231009064245.3573397-2-ruanjinjie@huawei.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

hid: cp2112: Fix IRQ shutdown stopping polling for all IRQs on chip

2023-11-20T10:59:22+00:00

[ Upstream commit dc3115e6c5d9863ec1a9ff1acf004ede93c34361 ]

Previously cp2112_gpio_irq_shutdown() always cancelled the
gpio_poll_worker, even if other IRQs were still active, and did not set
the gpio_poll flag to false. This resulted in any call to _shutdown()
resulting in interrupts no longer functioning on the chip until a
_remove() occurred (a.e. the cp2112 is unplugged or system rebooted).

Only cancel polling if all IRQs are disabled/masked, and correctly set
the gpio_poll flag, allowing polling to restart when an interrupt is
next enabled.

Signed-off-by: Danny Kaehn 
Fixes: 13de9cca514e ("HID: cp2112: add IRQ chip handling")
Link: https://lore.kernel.org/r/20231011182317.1053344-1-danny.kaehn@plexus.com
Signed-off-by: Benjamin Tissoires 
Signed-off-by: Sasha Levin

hid: cp2112: Fix duplicate workqueue initialization

2023-11-20T10:59:20+00:00

[ Upstream commit e3c2d2d144c082dd71596953193adf9891491f42 ]

Previously the cp2112 driver called INIT_DELAYED_WORK within
cp2112_gpio_irq_startup, resulting in duplicate initilizations of the
workqueue on subsequent IRQ startups following an initial request. This
resulted in a warning in set_work_data in workqueue.c, as well as a rare
NULL dereference within process_one_work in workqueue.c.

Initialize the workqueue within _probe instead.

Fixes: 13de9cca514e ("HID: cp2112: add IRQ chip handling")
Signed-off-by: Danny Kaehn 
Signed-off-by: Jiri Kosina 
Signed-off-by: Sasha Levin

HID: logitech-hidpp: Fix kernel crash on receiver USB disconnect

2023-10-06T14:09:14+00:00

hidpp_connect_event() has *four* time-of-check vs time-of-use (TOCTOU)
races when it races with itself.

hidpp_connect_event() primarily runs from a workqueue but it also runs
on probe() and if a "device-connected" packet is received by the hw
when the thread running hidpp_connect_event() from probe() is waiting on
the hw, then a second thread running hidpp_connect_event() will be
started from the workqueue.

This opens the following races (note the below code is simplified):

1. Retrieving + printing the protocol (harmless race):

	if (!hidpp->protocol_major) {
		hidpp_root_get_protocol_version()
		hidpp->protocol_major = response.rap.params[0];
	}

We can actually see this race hit in the dmesg in the abrt output
attached to rhbz#2227968:

[ 3064.624215] logitech-hidpp-device 0003:046D:4071.0049: HID++ 4.5 device connected.
[ 3064.658184] logitech-hidpp-device 0003:046D:4071.0049: HID++ 4.5 device connected.

Testing with extra logging added has shown that after this the 2 threads
take turn grabbing the hw access mutex (send_mutex) so they ping-pong
through all the other TOCTOU cases managing to hit all of them:

2. Updating the name to the HIDPP name (harmless race):

	if (hidpp->name == hdev->name) {
		...
		hidpp->name = new_name;
	}

3. Initializing the power_supply class for the battery (problematic!):

hidpp_initialize_battery()
{
        if (hidpp->battery.ps)
                return 0;

	probe_battery(); /* Blocks, threads take turns executing this */

	hidpp->battery.desc.properties =
		devm_kmemdup(dev, hidpp_battery_props, cnt, GFP_KERNEL);

	hidpp->battery.ps =
		devm_power_supply_register(&hidpp->hid_dev->dev,
					   &hidpp->battery.desc, cfg);
}

4. Creating delayed input_device (potentially problematic):

	if (hidpp->delayed_input)
		return;

	hidpp->delayed_input = hidpp_allocate_input(hdev);

The really big problem here is 3. Hitting the race leads to the following
sequence:

	hidpp->battery.desc.properties =
		devm_kmemdup(dev, hidpp_battery_props, cnt, GFP_KERNEL);

	hidpp->battery.ps =
		devm_power_supply_register(&hidpp->hid_dev->dev,
					   &hidpp->battery.desc, cfg);

	...

	hidpp->battery.desc.properties =
		devm_kmemdup(dev, hidpp_battery_props, cnt, GFP_KERNEL);

	hidpp->battery.ps =
		devm_power_supply_register(&hidpp->hid_dev->dev,
					   &hidpp->battery.desc, cfg);

So now we have registered 2 power supplies for the same battery,
which looks a bit weird from userspace's pov but this is not even
the really big problem.

Notice how:

1. This is all devm-maganaged
2. The hidpp->battery.desc struct is shared between the 2 power supplies
3. hidpp->battery.desc.properties points to the result from the second
   devm_kmemdup()

This causes a use after free scenario on USB disconnect of the receiver:
1. The last registered power supply class device gets unregistered
2. The memory from the last devm_kmemdup() call gets freed,
   hidpp->battery.desc.properties now points to freed memory
3. The first registered power supply class device gets unregistered,
   this involves sending a remove uevent to userspace which invokes
   power_supply_uevent() to fill the uevent data
4. power_supply_uevent() uses hidpp->battery.desc.properties which
   now points to freed memory leading to backtraces like this one:

Sep 22 20:01:35 eric kernel: BUG: unable to handle page fault for address: ffffb2140e017f08
...
Sep 22 20:01:35 eric kernel: Workqueue: usb_hub_wq hub_event
Sep 22 20:01:35 eric kernel: RIP: 0010:power_supply_uevent+0xee/0x1d0
...
Sep 22 20:01:35 eric kernel:  ? asm_exc_page_fault+0x26/0x30
Sep 22 20:01:35 eric kernel:  ? power_supply_uevent+0xee/0x1d0
Sep 22 20:01:35 eric kernel:  ? power_supply_uevent+0x10d/0x1d0
Sep 22 20:01:35 eric kernel:  dev_uevent+0x10f/0x2d0
Sep 22 20:01:35 eric kernel:  kobject_uevent_env+0x291/0x680
Sep 22 20:01:35 eric kernel:  power_supply_unregister+0x8e/0xa0
Sep 22 20:01:35 eric kernel:  release_nodes+0x3d/0xb0
Sep 22 20:01:35 eric kernel:  devres_release_group+0xfc/0x130
Sep 22 20:01:35 eric kernel:  hid_device_remove+0x56/0xa0
Sep 22 20:01:35 eric kernel:  device_release_driver_internal+0x19f/0x200
Sep 22 20:01:35 eric kernel:  bus_remove_device+0xc6/0x130
Sep 22 20:01:35 eric kernel:  device_del+0x15c/0x3f0
Sep 22 20:01:35 eric kernel:  ? __queue_work+0x1df/0x440
Sep 22 20:01:35 eric kernel:  hid_destroy_device+0x4b/0x60
Sep 22 20:01:35 eric kernel:  logi_dj_remove+0x9a/0x100 [hid_logitech_dj 5c91534a0ead2b65e04dd799a0437e3b99b21bc4]
Sep 22 20:01:35 eric kernel:  hid_device_remove+0x44/0xa0
Sep 22 20:01:35 eric kernel:  device_release_driver_internal+0x19f/0x200
Sep 22 20:01:35 eric kernel:  bus_remove_device+0xc6/0x130
Sep 22 20:01:35 eric kernel:  device_del+0x15c/0x3f0
Sep 22 20:01:35 eric kernel:  ? __queue_work+0x1df/0x440
Sep 22 20:01:35 eric kernel:  hid_destroy_device+0x4b/0x60
Sep 22 20:01:35 eric kernel:  usbhid_disconnect+0x47/0x60 [usbhid 727dcc1c0b94e6b4418727a468398ac3bca492f3]
Sep 22 20:01:35 eric kernel:  usb_unbind_interface+0x90/0x270
Sep 22 20:01:35 eric kernel:  device_release_driver_internal+0x19f/0x200
Sep 22 20:01:35 eric kernel:  bus_remove_device+0xc6/0x130
Sep 22 20:01:35 eric kernel:  device_del+0x15c/0x3f0
Sep 22 20:01:35 eric kernel:  ? kobject_put+0xa0/0x1d0
Sep 22 20:01:35 eric kernel:  usb_disable_device+0xcd/0x1e0
Sep 22 20:01:35 eric kernel:  usb_disconnect+0xde/0x2c0
Sep 22 20:01:35 eric kernel:  usb_disconnect+0xc3/0x2c0
Sep 22 20:01:35 eric kernel:  hub_event+0xe80/0x1c10

There have been quite a few bug reports (see Link tags) about this crash.

Fix all the TOCTOU issues, including the really bad power-supply related
system crash on USB disconnect, by making probe() use the workqueue for
running hidpp_connect_event() too, so that it can never run more then once.

Link: https://bugzilla.redhat.com/show_bug.cgi?id=2227221
Link: https://bugzilla.redhat.com/show_bug.cgi?id=2227968
Link: https://bugzilla.redhat.com/show_bug.cgi?id=2227968
Link: https://bugzilla.redhat.com/show_bug.cgi?id=2242189
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217412#c58
Cc: stable@vger.kernel.org
Signed-off-by: Hans de Goede 
Link: https://lore.kernel.org/r/20231005182638.3776-1-hdegoede@redhat.com
Signed-off-by: Benjamin Tissoires

HID: i2c-hid: fix handling of unpopulated devices

2023-10-06T07:14:19+00:00

A recent commit reordered probe so that the interrupt line is now
requested before making sure that the device exists.

This breaks machines like the Lenovo ThinkPad X13s which rely on the
HID driver to probe second-source devices and only register the variant
that is actually populated. Specifically, the interrupt line may now
already be (temporarily) claimed when doing asynchronous probing of the
touchpad:

	genirq: Flags mismatch irq 191. 00082008 (hid-over-i2c) vs. 00082008 (hid-over-i2c)
	i2c_hid_of 21-0015: Could not register for hid-over-i2c interrupt, irq = 191, ret = -16
	i2c_hid_of: probe of 21-0015 failed with error -16

Fix this by restoring the old behaviour of first making sure the device
exists before requesting the interrupt line.

Note that something like this should probably be implemented also for
"panel followers", whose actual probe is currently effectively deferred
until the DRM panel is probed (e.g. by powering down the device after
making sure it exists and only then register it as a follower).

Fixes: 675cd877c952 ("HID: i2c-hid: Rearrange probe() to power things up later")
Cc: Douglas Anderson 
Cc: Maxime Ripard 
Signed-off-by: Johan Hovold 
Tested-by: Dennis Gilmore 
Reviewed-by: Douglas Anderson 
Tested-by: Douglas Anderson 
Link: https://lore.kernel.org/r/20231002155857.24584-1-johan+linaro@kernel.org
Signed-off-by: Benjamin Tissoires

HID: Add quirk to ignore the touchscreen battery on HP ENVY 15-eu0556ng

2023-10-05T10:50:35+00:00

Like various other devices using similar hardware, this model reports a
perpetually empty battery (0-1%).

Join the others and apply HID_BATTERY_QUIRK_IGNORE.

Signed-off-by: Fabian Vogt 
Signed-off-by: Jiri Kosina